This proposal seeks to delineate the mechanisms underlying how movement information is represented by, and transferred between, the dominant and non-dominant arm controllers. Specifically, the proposed studies will determine whether information storage for the dominant and non-dominant arm systems shares the same working memory resources (Aim 1), test the hypothesis that interlimb transfer of visuomotor transformations is based on the development of an internal model of the transformations (Aim 2), and compare interlimb transfer of learning novel inertial loads to that of novel visuomotor transformations (Aim 3). These aims will be investigated by comparing the effects of initial training with one arm on subsequent performance with the other arm, in adaptation to altered visual displays of hand position and/or to novel inertial conditions. The information derived from these findings should prove essential for clinical rehabilitation applications that address motor learning in patients who must adapt to altered musculoskeletal conditions, as imposed by neural and orthopedic dysfunction. This information will be especially beneficial to those applications involving patients with unilateral damage, who are often forced to use the non-dominant arm for functional activities. Our findings on bimanual transfer of learning may lead to new insights that can help structure specific training protocols using opposite arm training to improve functioning in patients with hemiparesis.